Expected Utility Bruno Salcedo Cornell University · Decision Theory · Spring 2017 1 / 42 motivating examples powerball 2 / 42 powerball 3 / 42 powerball 4 / 42 st. petersburg paradox • Flip a fair coin until it lands tails • If we flipped the coin n times, you get $2n • How much would you be willing to pay to participate? 1 1 1 ∞ 1 E[2n ]= · 2+ · 4+ · 8+ . = · 2n = ∞ 2 4 8 2n n=1 X ∞ 1 ∞ 1 E[ log(2n)]= · log(2n) = log(2) · n = 2log(2) ≈ 0.60 2n 2n n=1 n=1 X X 5 / 42 von Neumann and Morgestern simple lotteries • A simple lottery is a tuple L =(p1, x1; p2, x2; ...pn, xn) – Monetary prizes x1,...,xn ∈ X ⊆ R – Probability distribution (p1,...,pn), pi is the probability of xi • Let L denote the set of simple lotteries • Example: L = (10, 0.2; 5, 0.1; 0, 0.3; −5, 0.4) b 10 0.2 0.1 b 5 bc 0.3 b 0 0.4 b −5 6 / 42 simplex • Simple lotteries given a fixed set of prizes x1,...,xn correspond to points in the n-dimensional simplex n n ∆ = (p1,...pn) ∈ R 0 ≤ pi ≤ 1 & p1 + . + pn =1 n o p3 b x1 0.1 (0, 0, 1) b L b L bc b x2 0.3 (0, 1, 0) b 0.6 (1, 0, 0) p2 b b x3 p1 7 / 42 simplex • Simple lotteries given a fixed set of prizes x1,...,xn correspond to points in the n-dimensional simplex n n ∆ = (p1,...pn) ∈ R 0 ≤ pi ≤ 1 & p1 + . + pn =1 n o b x1 0.1 (0, 0, 1) b L L bc b x2 0.3 (1, 0, 0) (0, 1, 0) 0.6 b x3 8 / 42 lottery mixtures • For 0 ≤ α ≤ 1 and lotteries L =(p1, x1; p2, x2; ...pn, xn) and M =(q1, x1; q2, x2; ...qn, xn) with the same set of prizes, define αL ⊕ (1 − α)M = r1, x1; r2, x2; ...rn, xn where rk = αpk + (1 − α)qk • Example: L = (0.5, 10; 0.5, 5), M = (0.8, 10; 0.2, 5), α =0.6 0.5 b 10 0.5 b 10 b 10 0.62 bc b 0.6 0.5 b 5 0.5 b 5 bc bc 0.8 b 10 0.8 b 10 bc 0.4 b 0.38 0.2 b 5 0.2 b 5 b 5 9 / 42 geometry of mixtures • With a fixed set of prizes x1,...,xn, mixtures between lotteries correspond to points in the line segment between them • The mixture weights determine the location within the segment (0, 0, 1) M b b 0.2L ⊕ 0.8M k M − L k .8 0 b L (1, 0, 0) (0, 1, 0) 10 / 42 lottery mixtures • Also possible to mix lotteries with different prizes • Example: L = (0.5, 10; 0.5, 5), M = (0.8, 20; 0.2, 5), α =0.6 0.5 b 10 0.5 b 10 b 10 0.3 bc b 0.6 0.5 b 5 0.5 b 5 bc bc b 20 0.32 0.8 b 20 0.8 b 20 bc 0.4 b 0.38 0.2 b 5 0.2 b 5 b 5 αL ⊕ (1 − α)M = (0.3, 10; 0.32, 20; 0.38, 5) 11 / 42 expected utility • Reported preferences ≻ on L • A utility function U : L → R for ≻ is an expected utility function if it can be written as n U(L)= pi u(xi ) Xk=1 for some function u : R → R • If you think of the prizes as a random variable x, then U(L)= EL [ u(x) ] • The function u is called a Bernoulli utility function 12 / 42 expected utility axioms • Axiom 1: (Preference order) ≻ is a asymmetric and negatively transitive • Axiom 2: (Continuity) For all simple lotteries L,M,N ∈ L, if L ≻ M ≻ N then there exist α, β ∈ (0, 1) such that αL ⊕ (1 − α)N ≻ M ≻ βN ⊕ (1 − β)N • Axiom 3: (Independence) For all lotteries L,M,N ∈ L and α ∈ (0, 1], if L ≻ M, then αL ⊕ (1 − α)N ≻ αM ⊕ (1 − α)N 13 / 42 continuity • The continuity axiom can be thought of as requiring that strict preference is preserved by sufficiently small perturbations in the probabilities – If L ≻ M, then so are lotteries which are close enough to L (hatched area) – This includes αL ⊕ (1 − α)N with α close enough to 1 (0, 0, 1) N b αL ⊕ (1 − α)N M b b b L (1, 0, 0) (0, 1, 0) 14 / 42 independence • If L is preferred to M, then a mixture of L with N is also preferred to a mixture of M with N using the same mixing weights • Independence gives the expected-utility structure • Similar to the independent-factors requirement from previous notes (expected utility is a form of additive separability) 15 / 42 example • How do you rank the following lotteries? 0.4 b 600 0.8 b 1500 L bc M bc 0.6 b 400 0.2 b −100 • How do you rank the following lotteries? b b 0.1 1500 0.5 1500 0.2 b 600 0.3 b 500 L′ bc M′ bc 0.3 b 500 0.1 b 400 0.4 b 400 0.1 b −100 • Independence says that if you prefer L to M, then you also prefer L′ to M′ • Note that L′ =0.5L ⊕ 0.5N and M′ =0.5M ⊕ 0.5N, for some lottery N (which lottery?) 16 / 42 allais’ paradox • How do you rank the following lotteries? b 0.1 5, 000, 000$ L bc b 1, 000, 000$ M bc b 1, 000, 000$ 1 1 1 0.89 0.01 b 0 • How do you rank the following lotteries? 0.11 b 1, 000, 000$ 0.1 b 5, 000, 000$ bc bc L2 M2 0.89 b 0 0.9 b 0 • Many people report L1 ≻ M1 and M2 ≻ L2 17 / 42 allais’ paradox • Note that we can write 1/11 b 0$ b 1 b b 1, 000, 000$ 0.11 0.11 b 10/11 5, 000, 000$ bc bc L1 M1 0.89 b 1 b 0.89 b 1 b 1, 000, 000$ 1, 000, 000$ 1/11 b 0$ b 1 b b 1, 000, 000$ 0.11 0.11 b 10/11 5, 000, 000$ bc bc L2 M2 0.89 b 1 b 0.89 b 1 b 0 0 • Independence would imply that L1 ≻ M1 if and only if L2 ≻ M2 (why?) 18 / 42 von neumann-morgenstern theorem Theorem: (a) A binary relation ≻ over L has an expected utility representation if and only if it satisfies axioms 1–3 (b) If U and V are expected utility representations of ≻, then there exist constants a,b ∈ R, a> 0, such that U( · )= a · V ( · )+ b 19 / 42 proof of necessity • Suppose U is an expected utility representation of ≻ • Axiom 1 follows from the same arguments as before • For 0 ≤ α ≤ 1 and lotteries L =(p1, x1; p2, x2; ...pn, xn) and M =(q1, x1; q2, x2; ...qn, xn) note that n V αL ⊕ (1 − α)M = αpi + (1 − α)qi · u(xi ) i=1 Xn = αpi u(xi )+(1 − α)qi u(xi ) i=1 Xn n = α pi u(xi )+(1 − α) (qi u(xi ) Xi=1 Xi=1 = αV (L)+(1 − α)V (M) • From here, it is straightforward to show that ≻ satisfies axioms 2 & 3 20 / 42 independence and linearity • Fix the set of prizes so that lotteries can be though of as vectors in ∆n • The following proposition that, under axioms 1–3, preferences are preserved under translations • This means that the indifference curves are parallel lines Proposition: Given lotteries L, M ∈ ∆n, and a vector N ∈ Rn, if L + N and M + N are also lotteries and ≻ satisfies axioms 1–3, then L ≻ M ⇔ (L + N) ≻ (M + N) 21 / 42 M + N b M b A b b L b L + N b B Proof sketch: • If (L + N) and (M + N) are lotteries, then so are A and B • A =0.5M ⊕ 0.5(L + N) and A =0.5L ⊕ 0.5(M + N) • Since A =0.5M ⊕ 0.5(L + N), independence says that if L ≻ M then B ≻ A • Since A =0.5L ⊕ 0.5(M + N), independence says that if B ≻ A then (L + N) ≻ (M + N) 22 / 42 risk aversion risk attitudes • For the rest of these slides, suppose u is strictly increasing (our decision maker always prefers more money) and twice continuously differentiable • Risk-neutral decision maker – E[ u(x)]= u(E[ x ]) for every random variable x • Risk-averse decision maker – E[ u(x) ] ≤ u(E[ x ]) for every r.v. x • Risk-loving decision maker – E[ u(x) ] ≥ u(E[ x ]) for every r.v. x 23 / 42 jensen’s inequality • A set is convex if it contains all the line-segments between its points • A function is concave if its hypograph is a convex set • Risk aversion is equivalent to u being concave u(x) b u(x2) u(E[ x ]) b E[ u(x)] b b u(x1) x x1 E[ x ] x2 24 / 42 certainty equivalent Definition: Given u, he certainty equivalent of a lottery x is the is the guaranteed amount of money that an individual with Bernoulli utility function u would view as equally desir- able as x, i.e., −1 CEu(x)= u (E[ u(x) ]) • Risk-neutral decision maker – CE(L)= E[ x ] for every r.v.